Image Forgery Detection

Finds edited or tampered regions in a real photo.

Slug	ID	Status	Tagged	Model version
`image-forgery-detection`	`5`	Active	Yes	`v2`

What it detects

Detects	Does not detect
Regions of a real image that were edited, such as splicing, copy-move or inpainting	Images that were AI-generated from scratch with no edited regions. Use `ai-image-detection` for those

Request

Slot	Accepted `contentType`
`original`	`image/jpeg`, `image/png`

{
  "taskTypeId": "5",
  "inputs": [
    { "slotName": "original", "fileName": "receipt.png", "contentType": "image/png", "sizeBytes": 832512 }
  ]
}

See Running a job for the full create, upload and finalize flow.

Limits

File size: up to the account upload limit, 10 MB by default.
Very large images may be analyzed at reduced resolution.

Result

Uses the standard result fields. Here "fake" means tampered: fakeConfidencePercent is the model's estimate that part of the image was edited.

{
  "isFake": true,
  "confidencePercent": 98.68,
  "fakeConfidencePercent": 98.68,
  "realConfidencePercent": 1.32,
  "tag": "fake",
  "tagBoundaries": { "real": 30, "likelyReal": 50, "likelyFake": 75 },
  "modelVersion": "v2"
}

Output artifacts

`kind`	File	Description
`result`	`result.json`	Full result with the raw tamper score
`heatmap`	`tampered-image-mask.png`	Pixel mask at the image's original resolution: white is untouched, black is tampered

Download them from artifacts on GET /jobs/{id}.

Cost and availability

Fixed price per job, set by your plan. The exact amount is returned as cost when you create the job. See Quotas & Credits.
Available on the cloud API; your plan decides which task types you can use.